Northwestern at #8, just behind Texas. Worth noting that he has Ball State at #9...and Michigan at #3, and IU at #5...
Phil Steele's "Most Improved Teams"
- Thread starter Sheffielder
- Start date
Not sure how you're defining "improved", but we were just fine in '09 with Kafka and '10 with Persa, two pretty inexperienced QBs heading into those seasons. Based on his results in the past, I still have confidence that McCall will develop one of the three QBs (or perhaps more than one) effectively this season.
For what it's worth, it was just a graphic I saw when he was a guest on College Football Live...they didn't spend too much time on the methodology. As a fan, I was honestly kind of upbeat to see us on the list since I'm admittedly going into the season a bit pessimistic (plan for the worst/hope for the best). I was really surprised to see Indiana.
I'm pretty certain Kafka and Persa had taken more snaps and had a good deal more game experience than either Thorson or Alviti at this point. Not that they weren't inexperienced, but Thorson has yet to take a snap in an actual game.
We were certainly competitive in those seasons, but each of those teams was worse than the year before, so "improved" is a stretch.
Going into the '10 season, Persa had had about as much game experience as Alviti did last year. In 2009, Persa played (poorly) against PSU when Kafka got hurt, and then he did OK the following week at Iowa. That was it for his experience heading into the '10 season, and no one really knew how good he'd be. I'd say the same thing about Kafka heading into '09. While he played a bit in '08 and did beat Minnesota with his running ability, no one knew how effective of a pocket passer he'd be.
Kafka was basically an unknown heading into '09, and he finished second team All-B1G. Persa was an unknown heading into '10, and he finished first team All-B1G with the coaches and second team All-B1G with the media. So yeah, that gives me some confidence that McCall will be able to develop the QBs this year. Sure, Thorson has zero collegiate game experience, but he's evidently a very talented and smart QB. Alviti and Oliver both had some decent game experience last year. I think one of them (or maybe more than one) will surprise us (in a good way) this year.
Last edited:
OK, so was the lack of improvement in each of those season the fault of the QBs? Again, Kafka was inexperienced heading into '09, and he ended up being second team All-B1G. Persa was inexperienced heading into '10, and even with the injury, he was first team All-B1G with the coaches. Was it really Kafka's fault that the 'Cats finished 8-4 in the '09 regular season after going 9-3 in '10? And do you blame Persa for the 'Cats then going 7-5 in the '10 regular season? They were 7-3 when he tore his Achilles, and I say it's very likely they would have finished the regular season at 8-4 if he had stayed healthy.
It's much hard to "improve" on a 9-3 or 8-4 record than it is on a 5-7 record. I'm just saying that I think the initial comment wasn't fair to Kafka or Persa, and it set the bar artificially low for this year's crop of QBs.
Last edited:
But we really did not have a healthy experienced QB for much of last year. So at worst, I see that as a draw. Or WR. We lost out top guy before the season even began and others along the way. Or RB. We went into last year expecting the return of atop RB and we know what happened with VM (though JJ eventually picked up the resulting slack) and Green seemed not himself. Or returner (again VM) Etc., Etc. Look at how out of sych we were at the beginning of last year. Just saying that overall, we look to be better going into this year than last even with an inexperienced QB. That said, we probably have a harder schedule.
You're on somewhat solid ground with Persa, who ran 83 plays in 2009 (although that is still 83 more than Thorson). Kafka, by his senior year, had thrown 145 passes in game conditions and had started several games, so I would maintain he had much more game experience than either Thorson or Alviti, not to mention better knowledge of the offensive system as he had played in three seasons to that point.
It is really about the time in practice that they work with the first team as well as actual game experience. Prior to the2008 season, Persa had had none. Prior to 2010, he had plenty.
Steel has a huge statistical archive and staff but part of his projection is based on returning to the mean. I think he sees NU as a team that should be between 5-7 and 8-4 under current management and since we have under performed, he expects the law of averages to work out. He picked us as most improved last year too so take it as with a grain of salt. One of the advantages of these preseason magazines is no one reds them after the season in over.
He is on pretty solid ground from a pre-season standpoint. By all accounts, our defense should be solid. We have a proven running back and a stable of additional backs that should keep him from being over-used over the course of the season. Therefore, if the O-Line improves, then the team as a whole should improve. If the o-line and QB play improve, then we will improve a lot. And if the O-line, QB, and WR play all improve... then watch out.
So, putting us on a list of potentially most improved teams is logical. We just have to wait and see if we can get the improvement at any and/or all those keep positions.
So, putting us on a list of potentially most improved teams is logical. We just have to wait and see if we can get the improvement at any and/or all those keep positions.
Big IFs. I think QB will be OK. WR depends on how healthy they are. OL? Huge question..................
But from the point of view of someone creating a "most improved" list, all three of these groups have lots of room to improve from last season. Just playing the odds that at least one will improve and the other two will not get worse is not a bad gamble at this point in the pre-season. Couple that with a solid defense and you get an improved team. For the purpose of this discussion we are not talking about the Rose Bowl... just "improved".
But like I said, significant improvement in all three of these areas and (as always) a lack of significant injuries, and this could be a very, very good team.
I agree. I think WR will be better, if they are healthy. I think QB will be OK and not the problem with NU's offense. I think the OL will be younger and weaker than last season. Overall I don't see a better record than 2014 because of the weak OL.
The big question for me is if the speed guys such as Vault, Shuler and Anderson and maybe a guy like Roberts stay healthy. Not being able to stretch the field last year really hurt us. Also, the receivers need to catch the ball.
I don't have Steele's yearbook yet, but as I recall the reversion to the mean concept (valid part of the story) is based on luck stats like net close losses and a rather complex set of power rankings to predict average performance. Think of it as moneyball without free agency. It does give some reason for optimism.
All we have to do is revert to the mean with injuries and we should improve. We have gotten killed there in the last couple years.
Hopefully the odds favor us for lesser injuries this season. The biggest swing factor other than injuries is the play of our QB, whoever that will be. I'm guessing it will be Thorsen.
The odds for injuries are the same every year. After flipping heads 100 times in a row, the odds of getting heads on the 101st flip are still 50/50. The odds don't change regardless of the outcome of previous trials. There is nothing to prevent us from having 80 consecutive years of horrible injuries. Regression to the mean is a matter of chance operating over many trials, the odds don't change for any given trial no matter how long one has had a string of good or bad luck.
I think we all know this, but I find it odd when people think "we're due for a change in luck" or a string of bad luck means our odds increase in our favor. That's how gamblers lose everything they have and then some.
I think we all know this, but I find it odd when people think "we're due for a change in luck" or a string of bad luck means our odds increase in our favor. That's how gamblers lose everything they have and then some.
Last edited:
We're all hoping for that. No matter how bad a string of luck has been, odds for any given year don't change, though. The probability of a given string of similar outcomes change. However, outcomes from year to year, like successive flips of a coin, are independent from one another. The probability of getting ten heads with ten consecutive coin flips is <0.001, but the probability of the next flip being heads is still 50/50. We all know this, but I'm tweaked when I hear "regression to the mean" as a reason for optimism. That's how gamblers lose everything.
To be pedantic, if a coin turns heads 100 times in a row, it must be rigged and I'm betting all my money on heads on the 101st flip.
I'm not betting on luck. I'm betting on them looking at the cause of the injuries and doing something about it to reduce the chances.
While it is true that last years results would have no effect on the number of injuries this year, it is also true that the expectarion is for a mean number of injuries. (Are we using our injury number to determine the mean or NCAA FB in general) If for example, last years number of injuries was one standard deviation above the mean, that would mean that there was more than an 80% chance that the number of injuries would be less this year. If something else is causing injuries to be higher in number(training, other issues) that is a different story. If you had 80 years of a larger number of injuries, for example, the mean would be higher or if it was using NCAA numbers, it would mean that you were doing something that could be corrected. And that kind of information tends to get around pretty fast
If the number of injuries in a particular year is one standard deviation above the mean (or more) is only 16% then the odds you will have that many injuries the following year is only 16%. It is not 50-50 as in a coin toss. Therefore, it would be likely we would have fewer injuries the following year. Doesn't mean we will, just that it is highly likely we will. Not because we are due but because the likelyhood of that many injuries is less than 16%. Now if our number of injuries was because of some outside factor, this does not apply.
Do you seriously think that I'm equating our chances of having the level of injuries we've had to the 50/50 odds of a coin flip? Of course not. The coin flip is simply an instructive analogy.
I don't know if our injury rate thhe last two years has even been one standard deviation above the mean. But assuming it has been, and using your 16% percentage for the area of a curve beyond +1 SD, yes the chances of having injuries at a rate of 1 standard deviation above the mean remains 16%, regardless of the fact that the chances of us having two seasons in a row (2013-2014) with injuries 1 SD above the mean is 2.56%. Each year is independent of other years (in theory). The point is there is no inevitability or "force" involved in reverting to the mean. Going into the 2014 season, we didn't think we'd have a higher injury rate than we had in 2013, but our injury rate was probably even worse last year.
In reality, our injuries last year will affect our performance this year because players are still recovering and have lost ground in their fitness and skill development compared with other Big Ten players. I think our likelihood of having performance problems because of past and in-season injuries is going to be higher than some percentage figure expected for our injury rates relative to other teams.
This is closer to throwing a single die and throwing a 6 time after time. The odds are the same each time. About 17% and last throw has no effect on the next one but the next one only has a probability of 17%
Coin flips, tossing dice...same thing. They are completely random and independent events. Don't get hung up on the coin flip analogy. There is nothing to prevent you from throwing a 6 ten times in a row, and the odds on the last toss remains the same as the first toss, about 17%.
My point is that when people mention or expect reversion to the mean, it sounds like they think the odds are somehow affected by the previous year's results.
I am just saying that the coin flip analogy indicates 50-50 chance while the die roll implies a 17% chance of landing on a particular number. If we were at the median last year, coin flip would be appropriate as 50% chance more injuries and 50% chance of fewer. But if our number of injuries was 1 standard deviation above the mean or more last year, then the die roll would be more appropriate and the likelyhood of a lower number of injuries would be about 84% and of a higher number no more than 16%. Just saying that if injuries are indeed random, there should be a strong probability of fewer injuries this year, not because of a revision to the mean or because of the last couple years but because the expected number of injuries is a significantly lower number than we saw.
I think he was speaking metaphorically, not literally, that they were both examples of chance.
Oh I agree, and I think your die analogy is better in that it is simple and more realistic probability-wise to our situation than my simple coin flip analogy. I just cringe when I hear "revert to the mean".
Even a 50/50 likelihood that comes up with more injuries than the mean will not be good for us. All depends on who gets injured.
Similar threads
